Accelerating artificial neurons with photonic circuits

Share this on social media:

Artistic representation of a neural network containing optically-interconnected Mach-Zehnder interferometers. The interferometer is the main component of the quantum memristor. (Credit: Equinox Graphics, University of Vienna)

Physicists at the University of Vienna have demonstrated a new device, called a quantum memristor, which may allow scientists to combine AI and quantum computing, in order to unlock unprecedented capabilities for research. 

The experiment, carried out in collaboration with the National Research Council (CNR) and the Politecnico di Milano in Italy, has been realised on an integrated quantum processor operating on single photons. The work has been recently published in the journal ‘Nature Photonics’.

Michele Spagnolo, who is the first author of the publication in the journal ‘Nature Photonics’ comments: ‘Unlocking the full potential of quantum resources within artificial intelligence is one of the greatest challenges of the current research in quantum physics and computer science. The group of Philip Walther of the University of Vienna has also recently demonstrated that robots can learn faster when using quantum resources and borrowing schemes from quantum computation. This new achievement represents one more step towards a future where quantum artificial intelligence becomes reality.

In recent years, artificial intelligence has become ubiquitous, with applications such as speech interpretation, image recognition, medical diagnosis, and many more. At the same time, quantum technology has been proven capable of computational power well beyond the reach of even the world’s largest supercomputer in certain applications. 

At the heart of artificial intelligence applications are mathematical models called neural networks. These models are inspired by the biological structure of the human brain, made of interconnected nodes. Just like our brain learns by constantly rearranging the connections between neurons, neural networks can be mathematically trained by tuning their internal structure until they become capable of human-level tasks: recognising our face, interpreting medical images for diagnosis, even driving our cars. Having integrated devices capable of performing the computations involved in neural networks quickly and efficiently has thus become a major research focus, both academic and industrial.

Memristors devices change their resistance depending on a memory of the past current, hence the name memory-resistor, or memristor. Immediately after its discovery, scientists realised that (among many other applications) the peculiar behaviour of memristors was surprisingly similar to that of neural synapses. The memristor has thus become a fundamental building block of neuromorphic architectures.

A group of experimental physicists from the University of Vienna, the National Research Council (CNR) and the Politecnico di Milano led by Professor Philip Walther and Dr Roberto Osellame, have now demonstrated that it is possible to engineer a device that has the same behaviour as a memristor, while acting on quantum states and being able to encode and transmit quantum information. In other words, a quantum memristor. Realizing such a device is challenging because the dynamics of a memristor tend to contradict the typical quantum behaviour.

By using single photons, i.e. single quantum particles of lights, and exploiting their unique ability to propagate simultaneously in a superposition of two or more paths, the physicists have overcome the challenge. In their experiment, single photons propagate along waveguides laser-written on a glass substrate and are guided on a superposition of several paths. 

One of these paths is used to measure the flux of photons going through the device and this quantity, through a complex electronic feedback scheme, modulates the transmission on the other output, thus achieving the desired memristive behaviour. Besides demonstrating the quantum memristor, the researchers have provided simulations showing that optical networks with quantum memristor can be used to learn on both classical and quantum tasks, hinting at the fact that the quantum memristor may be the missing link between artificial intelligence and quantum computing.


The space sector offers numerous oppurtunities for photonics technologies in the development and offering of satellite-based services. (Image: Shutterstock/Dima Zel)

14 June 2022

GasSight is a low-cost imager based on quantum technology that can produce real time video of methane gas. (Credit: QuantIC)

21 April 2022

The new processor offers computing density several orders higher than that of conventional electronic chips. (Image: University of Oxford/June Sang Lee)

15 June 2022

The space sector offers numerous oppurtunities for photonics technologies in the development and offering of satellite-based services. (Image: Shutterstock/Dima Zel)

14 June 2022

Preliminary testing of the quantum gravity gradiometer designed by Michael Holynski and colleagues at the University of Birmingham, which has now been shown to locate an underground tunnel with a positional accuracy of 20cm. Credit: Crown Copyright

13 June 2022